The Pathway from Research to STEM Careers

Candra Kou, Crystal Tran, Aaeena Joshi

Undergraduate Research

The image shows a blurred chin and torso of an individual with a white lab coat and blue gloves in the background with the focus being on one hand holding a pipette and another holding a microcentrifugal tubes. On the table, there is a container that is holding several microcentrifugal tubes

Scientist Pipetting into Microcentrifugal Tubes

Photo by Julia Koblitz on Unsplash

Undergraduate Research and Careers

  • Undergraduate research refers to mentored research that undergraduate students participate in conducting. Students participate in research to contribute to advance the field as well as gain experience to progress in their academic or professional career.
  • The correlation between undergraduate research to academic and professional career progression, the data from the study about how “Undergraduate Research Experiences Broaden Diversity in the Scientific Workforce” sought to analyze the correlation.

STUDY #1 :Research Participation

Does participating in any mentored research (low- or high-intensity) raise the probability of:

  • accepting a place in a science graduate programs
  • entering a STEM career

Is there a difference in results for low-intensity versus high-intensity experiences?

Explanation of data: Study #1

  • Study by Oxford University: “Undergraduate research experiences broaden diversity in the scientific workforce”

  • Variables Explored:

  • MRL_1d: 1 semester, low-intensity research
  • MRL_2d: 2 semesters, low-intensity research
  • MRH_1d: 1 semester, high-intensity research
  • MRH_2d: 2 semesters, high-intensity research
  • SciGradAccept: Accepted into a science grad program
  • STEM_Career: Currently in a STEM career

RESEARCH INTENSITY vs GRADUATE PROGRAM ACCEPTANCE

Heat map displaying the proportion of different research durations and intensities that got accepted into Science Grad Programs. High intensity research, regardless of duration (1 or 2 semesters) has equal proportion of acceptance into grad programs (37%). Low intensity research for 1 semester had the lowest proportion of acceptance into grad programs (28%), while low intensity research for 2 semesters had the highest proportion of acceptance into grad programs (43%). This indicates that higher intensity and duration of undergraduate research increases the likelihood of acceptance into science graduate programs.

Heat map displaying the proportion of different research duration and intensities that got students accepted into Science Grad Programs.

INTENSITY vs STEM CAREER

4 animated column charts for each intensity and duration of undergraduate research (high, low, 1 semester, 2 semesters), comparing career outcome (STEM or non-STEM careers) to the number of students in each.

LOW- vs HIGH-INTENSITY OUTCOMES

Bar graphs displaying the proportion of students who pursued STEM careers and who got accepted into scientific graduate school programs. The groups of students are divided by intensity and duration of undergraduate research. In both scientific grad school acceptance and engagement in STEM careers, having a longer duration of research (2 semesters) increases the likelihood of either case, regardless of research intensity. Notably, the proportion of high intensity research is greater than low intensity for engagement in STEM careers.

Bar graphs displaying the proportion of students who pursued STEM careers and who got accepted into scientific graduate school programs. The groups of students are divided by intensity and duration of undergraduate research.

CONCLUSION

  • From the visualizations created through this study, we observed that overall, longer research seems to be correlated to a greater rate of students pursuing a future STEM career and being accepted into graduate school for their career.
  • As for our second question, we found that the difference between outcomes of low-intensity and high-intensity was too varied to be conclusive of any trends, and as such concluded that the duration of research had the biggest impact.

STUDY #2: SOIL SCIENCE

  • Does participating in any mentored research (low- or high-intensity) raise the probability of: Does constitutional support increase the probability of availability of specialized programs in Universities?
  • Institutional support: Land Grants
  • Specialized Program: Soil Sciences
  • Is there a difference in results for institutions with vs. without land grant status?

Explanation of data: Study #2

  • Data collected by Soil Science Society of America Journal
  • Dataset name: “Data on universities offering undergraduate degrees that train students for soil science careers at universities in the USA and its territories”
  • Variables explored:
  • (Land Grant) - Indicates whether the institution is a land-grant university
  • (Soil Science) - Indicates if the institution offers soil science

LAND GRANT vs SOIL SCIENCE OPPERTUNITIES

Bar graph displaying the proportion of institutions with soil science programs from institutions who either do or do not receive land grants. Notably, institutions that do have land grant status are twice as likely to have Soil Science programs, compared to institutions without land grant status.

Bar graph displaying the proportion of institutions with soil science programs from institutions who either do or do not receive land grants.

CONCLUSION

  • From this second study, we concluded that availability of land grants was positively related to the proportion of available opportunities for certain STEM programs (specifically Soil Science in this case)
  • This means that institutions with land grants (and thus more funding) were able to provide more resources for students in the program

DATA SOURCE

  • Hernandez, P. R., Woodcock, A., Estrada, M., & Schultz, P. W. (2017). Undergraduate research experiences broaden diversity in the scientific workforce. BioScience, 68(3), 204–211. https://doi.org/10.1093/biosci/bix163
  • Brevik, E. C. (2019). Bachelors-Level Soil Science training at Land-Grant institutions in the United States and its territories. Natural Sciences Education, 48(1), 180021. https://doi.org/10.4195/nse2018.12.0021

RESOURCES #1

  • CRAN Packages By Name. (2017). R-Project.org. https://cran.r-project.org/web/packages/available_packages_by_name.html
  • Coder, R. (2024, January 5). Heat map in ggplot2. R CHARTS | a Collection of Charts and Graphs Made With the R Programming Language. https://r-charts.com/correlation/heat-map-ggplot2/
  • Holtz, Y. (n.d.). Animated barplot transition with R. https://r-graph-gallery.com/288-animated-barplot-transition.html

RESOURCES #2

  • A grammar of animated graphics. (n.d.). https://gganimate.com/index.html
  • Package index. (n.d.). https://gganimate.com/reference/index.html
  • Renderers provided by gganimate — renderers. (n.d.). https://gganimate.com/reference/renderers.html
  • Dodge overlapping objects side-to-side — position_dodge. (n.d.). https://ggplot2.tidyverse.org/reference/position_dodge.html

RESOURCES #3

  • LiquidBrain Bioinformatics. (2020, April 25). How to use gganimate in R | A RStudio Tutorial for Beginners [Video]. YouTube. https://www.youtube.com/watch?v=ccuZYgcwusU
  • Transition between several distinct stages of the data — transition_states. (n.d.). https://gganimate.com/reference/transition_states.html
  • Chang, W. (2025, June 9). 3.9 Adding Labels to a Bar Graph | R Graphics Cookbook, 2nd edition. https://r-graphics.org/RECIPE-BAR-GRAPH-LABELS.html

RESOURCES #4

  • Chang, W. (2025a, June 9). 3.2 Grouping Bars Together | R Graphics Cookbook, 2nd edition. https://r-graphics.org/RECIPE-BAR-GRAPH-GROUPED-BAR.html

NOTES AND COMMENTS

Co-authors: Candra Kou, Crystal Tran

2 Basic tools used: filter, pivot_longer

1 Intermediate tool used: Created an animated graph

Link to data set 1: https://datadryad.org/dataset/doi:10.5061/dryad.50m50

Link to data set 2: https://datadryad.org/dataset/doi:10.5061/dryad.qjq2bvqdj

Outside Source 1: https://academic.oup.com/bioscience/article/68/3/204/4831122?login=true

Outside Source 2: https://acsess.onlinelibrary.wiley.com/doi/10.1002/saj2.20140